forked from ray-project/ray
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathrelease_tests.yaml
6942 lines (5540 loc) · 149 KB
/
release_tests.yaml
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
# Global release test configuration file.
# All your release test configuration should go here. Adding release tests here
# will automatically enable them in the Buildkite release testing schedules
# (except they have frequency: manual).
# Here is an example configuration for reference:
#- name: example_test
# # Tests with the same group will be grouped in the Buildkite UI
# group: Example group
# # Provide the working directory which will be uploaded to the cluster
# working_dir: example_dir
#
# # How often to run the tests.
# # One of [manual, any, multi, nightly, nightly-3x, weekly].
# # Descriptions of each frequency (that's not immediately obvious):
# # - manual: Not run on a schedule, but can be manually run through the buildkite UI.
# # - nightly-3x: Run 3 times a week (Monday, Wednesday, Friday).
# frequency: weekly
# # Owning team. This field will be persisted to the database
# team: ml
#
# # Python version. This optional field determines which Python version to run tests
# # on. This must be a string!
# python: "3.7"
#
# # Cluster information
# cluster:
# # Location of cluster compute, relative to working_dir
# cluster_compute: cluster_compute.yaml
# # Autosuspend parameter passed to the cluster.
# # The cluster will automatically terminate if inactive for this
# # many minutes. Defaults to 10 if not set.
# autosuspend_mins: 10
# # Optional cloud_id to use instead of the default cloud
# cloud_id: cld_12345678
# # Alternatively, you can specify a cloud name
# cloud_name: anyscale_default_cloud
#
# # Run configuration for the test
# run:
# # If you want to wait for nodes to be ready, you can specify this here:
# wait_for_nodes:
# # Number of nodes
# num_nodes: 16
# # Timeout for waiting for nodes. If nodes are not up by then, the
# # test will fail.
# timeout: 600
#
# # Optional prepare script to be run on the cluster before the test script
# prepare: python prepare.py
# # The prepare command can have a separate timeout
# prepare_timeout: 300
#
# # Main script to run as the test script
# script: python workloads/train_small.py
# # Timeout in seconds. After this time the test is considered as failed.
# timeout: 600
#
# # You can specify smoke test definitions here. If a smoke test is triggered,
# # it will deep update the main test configuration with the values provided
# # here. Smoke tests will automatically run with IS_SMOKE_TEST=1 as en
# # environment variable and receive the --smoke-test flag as a parameter in the
# # run script.
# smoke_test:
# # Smoke tests can have different frequencies. A smoke test is only triggered
# # when the regular test is not matched.
# frequency: nightly
# # Here we adjust the run timeout down and run on less nodes. The test script
# # remains the same.
# run:
# timeout: 300
# wait_for_nodes:
# num_nodes: 4
# timeout: 600
#
# # After the test finished, this handler (in alerts/) will process the results.
# # It can then let the test fail, e.g. if a metric regression is observed.
# alert: default
#######################
# Cluster scaling tests
#######################
- name: cluster_tune_scale_up_down
group: Cluster tests
working_dir: cluster_tests
frequency: nightly
team: ml
cluster:
byod: {}
cluster_compute: cpt_autoscaling_1-3_aws.yaml
run:
timeout: 3600
script: RAY_AIR_NEW_PERSISTENCE_MODE=1 python workloads/tune_scale_up_down.py
wait_for_nodes:
num_nodes: 0
variations:
- __suffix__: aws
- __suffix__: gce
env: gce
frequency: manual
cluster:
cluster_compute: cpt_autoscaling_1-3_gce.yaml
alert: default
############################
# Batch Inference Benchmarks
############################
# 10 GB image classification raw images with 1 GPU.
# 1 g4dn.4xlarge
- name: torch_batch_inference_1_gpu_10gb_raw
group: AIR tests
working_dir: air_tests/air_benchmarks
frequency: nightly
team: data
cluster:
byod:
type: gpu
cluster_compute: compute_gpu_1_cpu_16_aws.yaml
run:
timeout: 500
script: python workloads/gpu_batch_inference.py --data-directory=10G-image-data-synthetic-raw --data-format raw
alert: default
variations:
- __suffix__: aws
- __suffix__: gce
env: gce
frequency: manual
cluster:
cluster_compute: compute_gpu_1_cpu_16_gce.yaml
# 10 GB image classification parquet with 1 GPU.
# 1 g4dn.4xlarge
- name: torch_batch_inference_1_gpu_10gb_parquet
group: AIR tests
working_dir: air_tests/air_benchmarks
frequency: nightly
team: data
cluster:
byod:
type: gpu
cluster_compute: compute_gpu_1_cpu_16_aws.yaml
run:
timeout: 500
script: python workloads/gpu_batch_inference.py --data-directory=10G-image-data-synthetic-raw-parquet --data-format parquet
alert: default
variations:
- __suffix__: aws
- __suffix__: gce
env: gce
frequency: manual
cluster:
cluster_compute: compute_gpu_1_cpu_16_gce.yaml
# 300 GB image classification raw images with 16 GPUs
# 4 g4dn.12xlarge
- name: torch_batch_inference_16_gpu_300gb_raw
group: AIR tests
working_dir: air_tests/air_benchmarks
frequency: nightly
team: data
python: "3.8"
cluster:
byod:
type: gpu
cluster_compute: compute_gpu_4x4_aws.yaml
run:
timeout: 1000
script: python workloads/gpu_batch_inference.py --data-directory 300G-image-data-synthetic-raw --data-format raw
wait_for_nodes:
num_nodes: 4
alert: default
variations:
- __suffix__: aws
- __suffix__: gce
env: gce
frequency: manual
cluster:
cluster_compute: compute_gpu_4x4_gce.yaml
# 300 GB image classification parquet data with 16 GPUs
# 4 g4dn.12xlarge
- name: torch_batch_inference_16_gpu_300gb_parquet
group: AIR tests
working_dir: air_tests/air_benchmarks
frequency: nightly
team: data
cluster:
byod:
type: gpu
cluster_compute: compute_gpu_4x4_aws.yaml
run:
timeout: 1000
script: python workloads/gpu_batch_inference.py --data-directory 300G-image-data-synthetic-raw-parquet --data-format parquet
wait_for_nodes:
num_nodes: 4
alert: default
variations:
- __suffix__: aws
- __suffix__: gce
env: gce
frequency: manual
cluster:
cluster_compute: compute_gpu_4x4_gce.yaml
# 10 TB image classification parquet data with heterogenous cluster
# 10 g4dn.12xlarge, 10 m5.16xlarge
- name: torch_batch_inference_hetero_10tb_parquet
group: AIR tests
working_dir: air_tests/air_benchmarks
frequency: weekly
team: data
cluster:
byod:
type: gpu
cluster_compute: compute_hetero_10x10_aws.yaml
run:
timeout: 2000
script: python workloads/gpu_batch_inference.py --data-directory 10T-image-data-synthetic-raw-parquet --data-format parquet
wait_for_nodes:
num_nodes: 20
alert: default
#########################
# AIR release tests
#########################
- name: tune_with_frequent_pausing
group: AIR tests
working_dir: air_tests
frequency: nightly-3x
team: ml
cluster:
byod:
runtime_env:
- RAY_memory_usage_threshold=0.5
- automatic_object_spilling_enabled=0
cluster_compute: frequent_pausing/compute_config_aws.yaml
run:
timeout: 600 # 10min
long_running: true
script: RAY_AIR_NEW_PERSISTENCE_MODE=1 python frequent_pausing/script.py
variations:
- __suffix__: aws
- __suffix__: gce
env: gce
frequency: manual
cluster:
cluster_compute: frequent_pausing/compute_config_gce.yaml
alert: default
- name: long_running_horovod_tune_test
group: AIR tests
working_dir: air_tests
frequency: weekly
team: ml
cluster:
byod:
type: gpu
post_build_script: byod_horovod_master_test.sh
cluster_compute: horovod/compute_tpl_aws.yaml
variations:
- __suffix__: aws
- __suffix__: gce
env: gce
frequency: manual
cluster:
cluster_compute: horovod/compute_tpl_gce.yaml
run:
timeout: 36000
script: RAY_AIR_NEW_PERSISTENCE_MODE=1 python horovod/workloads/horovod_tune_test.py
long_running: true
wait_for_nodes:
num_nodes: 2
smoke_test:
frequency: manual
run:
timeout: 3600
alert: default
- name: air_benchmark_data_bulk_ingest
group: AIR tests
working_dir: air_tests/air_benchmarks
frequency: nightly
team: ml
cluster:
byod:
type: gpu
cluster_compute: compute_data_20_nodes_aws.yaml
run:
timeout: 3600
script: RAY_AIR_NEW_PERSISTENCE_MODE=1 python workloads/data_benchmark.py --dataset-size-gb=200 --num-workers=20
wait_for_nodes:
num_nodes: 20
variations:
- __suffix__: aws
- __suffix__: gce
env: gce
frequency: manual
cluster:
cluster_compute: compute_data_20_nodes_gce.yaml
alert: default
# AIR benchmarks for XGBoost CUJ
- name: air_benchmark_xgboost_cpu_10
group: AIR tests
working_dir: air_tests/air_benchmarks
frequency: nightly
team: ml
cluster:
byod:
type: gpu
post_build_script: byod_xgboost_cuj_test.sh
cluster_compute: compute_xgboost_aws.yaml
run:
timeout: 36000
script: python workloads/xgboost_benchmark.py
wait_for_nodes:
num_nodes: 11
variations:
- __suffix__: aws
- __suffix__: gce
env: gce
frequency: manual
cluster:
cluster_compute: compute_xgboost_gce.yaml
smoke_test:
frequency: manual
run:
timeout: 1800
alert: default
# Ray AIR distributed Torch benchmarks
- name: air_benchmark_torch_mnist_cpu_4x1
group: AIR tests
working_dir: air_tests/air_benchmarks
frequency: nightly
team: ml
cluster:
byod:
type: gpu
cluster_compute: compute_cpu_4_aws.yaml
run:
timeout: 3600
script: RAY_AIR_NEW_PERSISTENCE_MODE=1 python workloads/torch_benchmark.py run --num-runs 3 --num-epochs 20 --num-workers 4 --cpus-per-worker 8
wait_for_nodes:
num_nodes: 4
variations:
- __suffix__: aws
- __suffix__: gce
env: gce
frequency: manual
cluster:
cluster_compute: compute_cpu_4_gce.yaml
alert: default
- name: air_benchmark_torch_mnist_gpu_4x4
group: AIR tests
working_dir: air_tests/air_benchmarks
frequency: weekly
team: ml
cluster:
byod:
type: gpu
cluster_compute: compute_gpu_4x4_aws.yaml
run:
timeout: 4800
script: RAY_AIR_NEW_PERSISTENCE_MODE=1 python workloads/torch_benchmark.py run --num-runs 3 --num-epochs 120 --num-workers 16 --cpus-per-worker 4 --batch-size 1024 --use-gpu
wait_for_nodes:
num_nodes: 4
smoke_test:
frequency: nightly
cluster:
cluster_compute: compute_gpu_2x2_aws.yaml
run:
timeout: 3600
script: RAY_AIR_NEW_PERSISTENCE_MODE=1 python workloads/torch_benchmark.py run --num-runs 3 --num-epochs 60 --num-workers 4 --cpus-per-worker 4 --batch-size 512 --use-gpu
wait_for_nodes:
num_nodes: 2
variations:
- __suffix__: aws
- __suffix__: gce
env: gce
frequency: manual
cluster:
cluster_compute: compute_gpu_4x4_gce.yaml
smoke_test:
frequency: manual
alert: default
- name: air_benchmark_torch_mnist_cpu_1x4
group: AIR tests
working_dir: air_tests/air_benchmarks
frequency: nightly
team: ml
cluster:
byod:
type: gpu
cluster_compute: compute_cpu_1_aws.yaml
run:
timeout: 3600
script: RAY_AIR_NEW_PERSISTENCE_MODE=1 python workloads/torch_benchmark.py run --num-runs 3 --num-epochs 20 --num-workers 4 --cpus-per-worker 2
variations:
- __suffix__: aws
- __suffix__: gce
env: gce
frequency: manual
cluster:
cluster_compute: compute_cpu_1_gce.yaml
alert: default
- name: air_benchmark_torch_mnist_cpu_4x4
group: AIR tests
working_dir: air_tests/air_benchmarks
frequency: nightly
team: ml
cluster:
byod:
type: gpu
cluster_compute: compute_cpu_4_aws.yaml
run:
timeout: 5400
script: RAY_AIR_NEW_PERSISTENCE_MODE=1 python workloads/torch_benchmark.py run --num-runs 3 --num-epochs 20 --num-workers 16 --cpus-per-worker 2
wait_for_nodes:
num_nodes: 4
variations:
- __suffix__: aws
- __suffix__: gce
env: gce
frequency: manual
cluster:
cluster_compute: compute_cpu_4_gce.yaml
alert: default
- name: air_benchmark_tune_torch_mnist
group: AIR tests
working_dir: air_tests/air_benchmarks
frequency: nightly
team: ml
cluster:
byod:
type: gpu
cluster_compute: compute_cpu_8_aws.yaml
run:
timeout: 3600
script: RAY_AIR_NEW_PERSISTENCE_MODE=1 python workloads/tune_torch_benchmark.py --num-runs 3 --num-trials 8 --num-workers 4
wait_for_nodes:
num_nodes: 8
variations:
- __suffix__: aws
- __suffix__: gce
env: gce
frequency: manual
cluster:
cluster_compute: compute_cpu_8_gce.yaml
alert: default
- name: air_benchmark_tune_torch_mnist_gpu
group: AIR tests
working_dir: air_tests/air_benchmarks
frequency: nightly
team: ml
cluster:
byod:
type: gpu
cluster_compute: compute_gpu_4x4_aws.yaml
run:
timeout: 3600
script: RAY_AIR_NEW_PERSISTENCE_MODE=1 python workloads/tune_torch_benchmark.py --num-runs 2 --num-trials 4 --num-workers 4 --use-gpu
wait_for_nodes:
num_nodes: 4
variations:
- __suffix__: aws
- __suffix__: gce
env: gce
frequency: manual
cluster:
cluster_compute: compute_gpu_4x4_gce.yaml
alert: default
# Ray AIR distributed Tensorflow benchmarks
- name: air_benchmark_tensorflow_mnist_cpu_4x1
group: AIR tests
working_dir: air_tests/air_benchmarks
frequency: nightly
team: ml
cluster:
byod:
type: gpu
cluster_compute: compute_cpu_4_aws.yaml
run:
timeout: 5400
script: RAY_AIR_NEW_PERSISTENCE_MODE=1 python workloads/tensorflow_benchmark.py run --num-runs 3 --num-epochs 20 --num-workers 4 --cpus-per-worker 8
wait_for_nodes:
num_nodes: 4
variations:
- __suffix__: aws
- __suffix__: gce
env: gce
frequency: manual
cluster:
cluster_compute: compute_cpu_4_gce.yaml
alert: default
- name: air_benchmark_tensorflow_mnist_cpu_1x4
group: AIR tests
working_dir: air_tests/air_benchmarks
frequency: nightly
team: ml
cluster:
byod:
type: gpu
cluster_compute: compute_cpu_1_aws.yaml
run:
timeout: 5400
script: RAY_AIR_NEW_PERSISTENCE_MODE=1 python workloads/tensorflow_benchmark.py run --num-runs 3 --num-epochs 20 --num-workers 4 --cpus-per-worker 2
variations:
- __suffix__: aws
- __suffix__: gce
env: gce
frequency: manual
cluster:
cluster_compute: compute_cpu_1_gce.yaml
alert: default
- name: air_benchmark_tensorflow_mnist_cpu_4x4
group: AIR tests
working_dir: air_tests/air_benchmarks
frequency: nightly
team: ml
stable: false
cluster:
byod:
type: gpu
cluster_compute: compute_cpu_4_aws.yaml
run:
timeout: 5400
script: RAY_AIR_NEW_PERSISTENCE_MODE=1 python workloads/tensorflow_benchmark.py run --num-runs 3 --num-epochs 20 --num-workers 16 --cpus-per-worker 2
wait_for_nodes:
num_nodes: 4
variations:
- __suffix__: aws
- __suffix__: gce
env: gce
frequency: manual
cluster:
cluster_compute: compute_cpu_4_gce.yaml
alert: default
- name: air_benchmark_tensorflow_mnist_gpu_4x4
group: AIR tests
working_dir: air_tests/air_benchmarks
frequency: weekly
team: ml
stable: false
cluster:
byod:
type: gpu
cluster_compute: compute_gpu_4x4_aws.yaml
run:
timeout: 5400
script: RAY_AIR_NEW_PERSISTENCE_MODE=1 python workloads/tensorflow_benchmark.py run --num-runs 3 --num-epochs 200 --num-workers 16 --cpus-per-worker 4 --batch-size 1024 --use-gpu
wait_for_nodes:
num_nodes: 4
smoke_test:
frequency: nightly
cluster:
cluster_compute: compute_gpu_2x2_aws.yaml
run:
script: RAY_AIR_NEW_PERSISTENCE_MODE=1 python workloads/tensorflow_benchmark.py run --num-runs 3 --num-epochs 60 --num-workers 4 --cpus-per-worker 4 --batch-size 512 --use-gpu
wait_for_nodes:
num_nodes: 2
variations:
- __suffix__: aws
- __suffix__: gce
env: gce
frequency: manual
cluster:
cluster_compute: compute_gpu_4x4_gce.yaml
smoke_test:
frequency: manual
alert: default
- name: air_benchmark_pytorch_training_e2e_gpu_1x1_20gb
group: AIR tests
working_dir: air_tests/air_benchmarks
frequency: nightly
team: ml
cluster:
byod:
type: gpu
cluster_compute: compute_gpu_1_aws.yaml
run:
timeout: 3600
script: RAY_AIR_NEW_PERSISTENCE_MODE=1 python workloads/pytorch_training_e2e.py --data-size-gb 20
alert: default
variations:
- __suffix__: aws
- __suffix__: gce
env: gce
frequency: manual
cluster:
cluster_compute: compute_gpu_1_gce.yaml
- name: air_benchmark_pytorch_training_e2e_gpu_4x4_100gb
group: AIR tests
working_dir: air_tests/air_benchmarks
frequency: nightly
team: ml
stable: false
cluster:
byod:
type: gpu
cluster_compute: compute_gpu_4x4_aws.yaml
run:
timeout: 10800
script: RAY_AIR_NEW_PERSISTENCE_MODE=1 python workloads/pytorch_training_e2e.py --data-size-gb=100 --num-workers=16
wait_for_nodes:
num_nodes: 4
alert: default
variations:
- __suffix__: aws
- __suffix__: gce
env: gce
frequency: manual
cluster:
cluster_compute: compute_gpu_4x4_gce.yaml
# Test tiny, and medium input files to check that performance stays about
# constant.
- name: ray-data-resnet50-ingest-file-size-benchmark
group: AIR tests
working_dir: air_tests/air_benchmarks/mlperf-train
stable: false
frequency: nightly
team: data
cluster:
byod:
type: gpu
runtime_env:
- RAY_task_oom_retries=50
- RAY_min_memory_free_bytes=1000000000
cluster_compute: compute_cpu_16.yaml
run:
timeout: 3600
script: bash file_size_benchmark.sh
variations:
- __suffix__: aws
- __suffix__: gce
env: gce
frequency: manual
cluster:
cluster_compute: compute_gce_cpu_16.yaml
# Test huge files to check that we do not OOM.
- name: ray-data-resnet50-ingest-out-of-memory-benchmark
group: AIR tests
working_dir: air_tests/air_benchmarks/mlperf-train
stable: false
frequency: nightly
team: data
cluster:
byod:
type: gpu
runtime_env:
- RAY_task_oom_retries=50
- RAY_min_memory_free_bytes=1000000000
cluster_compute: compute_cpu_16.yaml
run:
timeout: 3600
script: bash oom_benchmark.sh
variations:
- __suffix__: aws
- __suffix__: gce
env: gce
frequency: manual
cluster:
cluster_compute: compute_gce_cpu_16.yaml
#######################
# AIR examples
#######################
# Test additional CPU nodes for preprocessing.
- name: air_example_dreambooth_finetuning
group: AIR examples
working_dir: air_examples/dreambooth
stable: false
frequency: weekly
team: ml
cluster:
byod:
type: gpu
cluster_compute: dreambooth_compute_aws.yaml
run:
timeout: 1800
script: pip install -Ur dreambooth/requirements.txt && RAY_AIR_NEW_PERSISTENCE_MODE=1 bash dreambooth_run.sh
artifact_path: /tmp/artifacts/example_out.jpg
# variations: A10G not available on GCE, yet.
- name: air_example_dreambooth_finetuning_lora
group: AIR examples
working_dir: air_examples/dreambooth
stable: false
frequency: weekly
team: ml
cluster:
byod:
type: gpu
cluster_compute: dreambooth_compute_aws.yaml
run:
timeout: 1800
script: pip install -Ur dreambooth/requirements.txt && bash dreambooth_run.sh --lora
artifact_path: /tmp/artifacts/example_out.jpg
- name: air_example_gptj_deepspeed_fine_tuning
group: AIR examples
working_dir: air_examples/gptj_deepspeed_finetuning
python: "3.9"
frequency: weekly
team: ml
cluster:
byod:
type: gpu
pip:
- myst-parser==0.15.2
- myst-nb==0.13.1
- jupytext==1.13.6
cluster_compute: gptj_deepspeed_compute_aws.yaml
run:
timeout: 4500
script: python test_myst_doc.py --path gptj_deepspeed_fine_tuning.ipynb
variations:
- __suffix__: aws
- __suffix__: gce
env: gce
frequency: manual
cluster:
cluster_compute: gptj_deepspeed_compute_gce.yaml
- name: air_example_dolly_v2_lightning_fsdp_finetuning
group: AIR examples
working_dir: air_examples/dolly_v2_lightning_fsdp_finetuning
frequency: weekly
team: ml
cluster:
byod:
type: gpu
pip:
- "datasets"
- "evaluate"
- "scikit-learn"
- "boto3"
- myst-parser==0.15.2
- myst-nb==0.13.1
- jupytext==1.13.6
post_build_script: byod_dolly_test.sh
cluster_compute: dolly_v2_fsdp_compute_aws.yaml
run:
timeout: 4700
script: python test_myst_doc.py --path lightning-llm-finetuning-7b.ipynb
# variations: TODO(jungong): add GCP variation.
- name: air_example_vicuna_13b_lightning_deepspeed_finetuning
group: AIR examples
working_dir: air_examples/vicuna_13b_lightning_deepspeed_finetuning
python: "3.8"
frequency: weekly
team: ml
cluster:
byod:
type: gpu
pip:
- myst-parser==0.15.2
- myst-nb==0.13.1
post_build_script: byod_vicuna_test.sh
cluster_compute: vicuna_13b_deepspeed_compute_aws.yaml
run:
timeout: 4700
script: python test_myst_doc.py --path vicuna_13b_lightning_deepspeed_finetune.ipynb
#####################################
# Workspace templates release tests #
#####################################
- name: workspace_template_batch_inference
group: Workspace templates
working_dir: workspace_templates/01_batch_inference
python: "3.9"
frequency: nightly-3x
team: data
cluster:
byod:
type: gpu
cluster_compute: ../testing/compute_configs/gpu/aws.yaml
run:
timeout: 600
script: jupyter nbconvert --to script --output _test start.ipynb && ipython _test.py
variations:
- __suffix__: aws
- __suffix__: gce
env: gce
frequency: manual
cluster:
cluster_compute: ../testing/compute_configs/gpu/gce.yaml
- name: workspace_template_many_model_training
group: Workspace templates
working_dir: workspace_templates/02_many_model_training
python: "3.9"
frequency: nightly-3x
team: ml
cluster:
byod:
type: gpu
cluster_compute: ../testing/compute_configs/cpu/aws.yaml
run:
timeout: 600
script: jupyter nbconvert --to script --output _test start.ipynb && RAY_AIR_NEW_PERSISTENCE_MODE=1 ipython _test.py
variations: